Barrier option valuation: Monte Carlo and historical simulations (Excel)
HTML-код
- Опубликовано: 20 июн 2022
- How one can value exotic options? The most straightforward method would be to utilise simulations. Today we are discussing Monte Carlo and historical bootstrap applications for barrier option valuation and analysing their applicability and limitations.
Don't forget to subscribe to NEDL and give this video a thumbs up for more videos in Derivatives!
Please consider supporting NEDL on Patreon: / nedleducation
You can find the spreadsheets for this video and some additional materials here: drive.google.com/drive/folders/1sP40IW0p0w5IETCgo464uhDFfdyR6rh7
Please consider supporting NEDL on Patreon: www.patreon.com/NEDLeducation
Wow Savva your a legend! Thanks for the video, I like how you include multiple models in your videos where possible and compare the pros and cons of each, that knowledge is very useful, thanks!
Yes me too :)
I love the proof of concept you designed with this video! Indicating the reasons for why we should select a non-normal distribution and how to spot it from the data set. I also didn't realize that barriers were a hard knock-out/in. I figured that was determined by observation frequency, but I guess the daily returns are the observation frequency. That definitely poses more risk than I originally anticipated with barrier options. I'm definitely looking forward to seeing the Python implementation of what you've demonstrated here with non-normal distributions. Also, love how you brought up the irrelevance of older data with time. A good reason to roll the distribution with the passage of time. As you touched on in your previous video, EWMA volatility sort of does this, but for volatility estimation obviously. Using a lambda decay factor of 0.975 should mean basically anything beyond around 100 days old will have a miniscule impact on the volatility calculation. From what I've done with EWMA and GARCH, EWMA is a bit overly sensitive whereas GARCH tends to not reach the volatility peaks that we often see. It's always interesting balancing the trade offs of relying on current data, historical data, or a weighted blend. The concept should be interesting when applied to distribution fitting and selection for a short period of time as options often are. I'm expecting to see more skew than with the long-term data driven distributions that are typically used.
What the heck, looks like someone decided to bot spam your video comment section.
To mention on observation frequency Savva used daily observation here but you can also use others like fortnightly where you only check the close price once every 2 weeks, you could change the selection of the min/max price indicators to reflect this instead of having all days of the sim results selected
@@King-Mac Thank you for the additional clarification
Sava good video as always but I am wonding whether you could create video on how to do 5 Year Rolling XIRR in Excel as I found it hard to do this in Excel.
Thanks a lot for you work. Can I use the same file for a currency option?
Hi, and glad you liked the video! Yes, absolutely, the application for currency options would be identical.
For the path formula could you use a t distribution or another non normal distribution in place of the Norm Dist?
Hi Shane, and thanks for the excellent question! Yes, of course, you could simulate any distribution on returns in Monte Carlo (this is one of the strengths of the approach).
@@NEDLeducation awesome news. Do you know the excel formula?
@@NEDLeducation also fantastic channel, really great content
Excellent video. However, please work on the technical issue of low sound. You are almost inaudible.
Low sound again. I can barely hear what you're saying.